I had a problem a while back where a mechanical relay controlled a motor and it caused emi that caused a bit of havoc on the controller. Changing the mechanical relay for a solid state relay solved the problem.

In another scenario I have the inductive load controlled by a mechanical switch which causes the same problem. Given that the voltage is 220v ac, would it not be possible to connect a suitably large capacitor across the switch like what is done on the controller or is this looking for trouble?

 

Looking for trouble. You need a TVS type solid-state suppressor, with maybe an RC network.

Things that will help: move the inductive load several meters from the controller. Put the controller in a metallic (shielded) enclosure. Feed only the control signal to the relay or other power switch with a twisted pair. Even better if the power switch is opto-isolated from the inductive load side(like the SSR you got), with no common power supply connection back to the controller(separate supplies for the controller and the load/switch with no common ground).

 

I had a problem a while back where a mechanical relay controlled a motor and it caused emi that caused a bit of havoc on the controller. Changing the mechanical relay for a solid state relay solved the problem.

This type of R/C snubber usually cures it for me on AC coils.
Max.

 

As has been mentioned consider a RC network.

Type Q/QRL (Quencharc) Arc Suppressor/Snubber Network.

Just as a few examples.

Ron

 

Commutation between Inductor and Capacitor is what you're after...XL=XC.

 

Commutation between Inductor and Capacitor is what you're after...XL=XC.

That can cause oscillations at the resonant LC frequency.
That's why all the practical snubbers also have a resistor to absorb the energy and damp the resonance.

 

As has been mentioned consider a RC network.

Type Q/QRL (Quencharc) Arc Suppressor/Snubber Network.

Just as a few examples.

Ron

Those are the same ones we used on a persistent problem just like the thread starter described.

We had 1/3 HP AC motors being switched my mechanical relays and had machines rebooting themselves quite a bit. Quencharc snubbers got us maybe 85-90% of the way there, but we still had an unacceptable failure rate. We switched to SSRs and that solved it.

I would say it's worth trying the snubber first and seeing if it's enough. If not, you could always make the switch a low power control circuit which signals an SSR. SSRs with AC voltage signal inputs are more expensive and less readily available than DC control models, so this idea might require a DC power supply. Alternately, if you stick with high voltage, it might be more expensive than one would expect to get the suitable SSR.

 

The original post was August 2015. I hope the thread starter solved his problem by now.

We had 1/3 HP AC motors being switched my mechanical relays and had machines rebooting themselves quite a bit. Quencharc snubbers got us maybe 85-90% of the way there, but we still had an unacceptable failure rate. We switched to SSRs and that solved it.

Been there and played that game. Exactly, with the reboots. Used snubbers and then went SSR which solved the problem.

Ron

 

The original post was August 2015. I hope the thread starter solved his problem by now.


Been there and played that game. Exactly, with the reboots. Used snubbers and then went SSR which solved the problem.

Ron

OMG! Didn't notice the dates. I saw new replies and just assumed it was a recent thread. Sometimes I wish this forum was a little better at discouraging necro-posts. Oh, well!

 

This is a immunity problem at your processor.
When in Europe check the CE_EMC report for this processor.
First of all you have to identify the real reason by measuring the processor power supply reset lines and inputs.
Are there open inputs if unused make then output or use a pull up resistor.
in pc power input should have low esr cap parallel ( eq 100nF+ 10uf+ 100uf) plus a dual coil transformer.
( transformer working like a long twisted wire)
The reset circuit should carry a simple network 100nF and resistor.
The tracks for input and power should be separated and as short as possible.


Picbuster